Art. XXI.—On the Languages Spoken Beyond the North Western Frontier

501

ART. XXI.—On the Languages spoken beyond the North- Western Frontier of India. With a map. By GEORGE A. GRIERSON, O.I.E., Ph.D., I.C.S.

BETWEEN the north-west frontier of our Indian Empire and the Pamirs there is a tract of mountainous country inhabited by many different nationalities, speaking many different tongues. The Pamirs themselves are a polyglot region. Taking Zebak, for instance, the district round it is the home of no less than four distinct speeches— one West-Iranian, Persian, and three East-Iranian, Wakhl, Shighni, and Ishkashaml. These last belong to the same Aryan group as Paksbto. To the south-east of the Pamirs we come to the Burushaski spoken in Hunza and Nagar, a language of Scythian stock, whose immediate affinities have not yet been identified. South-east, again, of the Burushaski area we come to Baltistan, where another Scythian language, the Tibeto-Burman Baltl, is the vernacular. In the valley of Kashmir, there is Kashmiri, and in the lower reaches of the Jhelum and in the Murree Hills, Chhibhall, both of which are Indo - Aryan, and can be traced up to ancient Sanskrit. West of the Chhibhall tract lies the British district of Hazara, of which the principal language is a form of Western PanjabI. Crossing the Indus we come to the Northern Pakhto dialect of Pakshto spoken in Peshawar, Swat, and Bajaur. West of Bajaur, beyond the Kunar River, we come to Laghman. North of Laghman lies Kafiristan, through which we again reach the Pamirs. We have now described a circle, and it remains to consider the interior portion of this tract. It consists of a number of river systems. The first is the Gilgit Valley, leading Downloaded from https://www.cambridge.org/core. University of Warwick, on 21 May 2018 at 04:09:30, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0035869X00027878 502 LANGUAGES SPOKEN BEYOND THE

into the Indus shortly after the latter debouches from Baltistan. Lower down, the Tanglr and the Kandia fall into the Indus, which has hitherto been flowing westwards, but takes a southerly course after its junction with the latter river. The language of the Gilgit Valley, and of the Indus Valley from Baltistan to the Tanglr, is Shlna in various dialects. This form of speech also extends to the south-east of the last-named river, and occupies a large block of mountain country between Baltistan and the valley of Kashmir. Prom its junction with the Kandia to its entry into British territory, the Indus runs in a southerly direction through groups of hills, known collectively as the Indus Kohistan, and inhabited by a number of wild tribes who all speak varieties of a language of Indo-Aryan origin, which, like Chhibhali, can be traced to ancient Sanskrit, and which is called Indus-KohistanI or Maiya.1 Colonel Biddulph has given us vocabularies of two of these dialects under the name of Gowro and Ohiliss. The Linguistic Survey of India, on which I am at present engaged, gives further details, including a brief grammar and specimens. To the west of the Indus-Kohistan lie, in order, the valleys of the Swat, the Panjkora, and the Kunar. Those of the first two are known as the Swat- and Panjkora-Kohistans respectively. Here the language of the bulk of the people was formerly an Indian one, allied to Indus-KohistanI, but is now, owing to Pathan domination, almost invariably Pakshto. Only a faithful few still cling to their ancient language, though they have abandoned their Aryan religion, and the dialects which they speak are called Garwl and Torwall. These three, Indus-KohistanI, Garwl, and Torwall, together form one well-defined group of languages, Indo-Aryan in origin, and evidently descended from ancient Sanskrit. They form a connecting link in the chain of North-Western Indo-Aryan languages, commencing with SindhI, and passing via Western PanjabI, through them, into Chhibhali and

1 The sign ~ over a vowel indicates a nasal pronunciation. Downloaded from https://www.cambridge.org/core. University of Warwick, on 21 May 2018 at 04:09:30, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0035869X00027878 NORTH-WESTERN FRONTIER OF INDIA. 503

Kashmiri. The Survey has made available grammars, vocabularies, and specimens of all of them. In this part of the country, Pakshto itself hardly gets further west than the hills forming the eastern side of the Kunar Valley. Nowhere does it cross that river. North of the Swat and Panjkora Valleys we find the country of Chitral, lying on both sides of the Kunar River, which is here known as the Qashqar, Chitrar, or (to Europeans) Chitral. The main speech of this country is called Chitrarl, or Kho-war, and is spoken as far east as Yasin, where it marches with BurushaskI and Shlna. Kho-war is evidently related to the latter language. They form a pair belonging to the Irano-Indian stock, and to the Indian branch of that family. They are hence to be classed as Indo-Aryan. This is at once established by a consideration of their phonetic systems, but their grammars present certain peculiarities which will be alluded to shortly. The two main affluents of the Chitral-Kunar River are the Bashgal and the Wai gal, both of which join it on the west after passing through the hill country of Kafiristan. The first-named is the most northern, and takes its rise in the southern face of the Hindu Kush. It joins the Chitral near the village of Narsat. The Waigal, after itself receiving the waters of the Wezgal, falls into the Chitral some way below Asmar. It is formed in the interior of Kafiristan. The Valley of the Bashgal is the home of the Bashgall language, which is the speech of the Siah-push Kafirs generally. A vocabulary and a few grammatical forms have been published by Colonel Biddulph, and a formal grammar by Colonel Davidson is now, I believe, in the press. East of the Bashgal Valley, Wasi-veri, another Kafir language, is spoken in the Wezgal Valley. A grammatical sketch, specimens, and a vocabulary of this will be published by the Linguistic Survey. It is evidently distantly related to Bashgall, and, like the remaining Kafir languages, is spoken by the Sufld-push Kafirs. These remaining ones are Ashkun, spoken in the heart of the Kafir country, and Wai, the language of the Waigal Valley. Regarding Downloaded from https://www.cambridge.org/core. University of Warwick, on 21 May 2018 at 04:09:30, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0035869X00027878 504 LANGUAGES SPOKEN BEYOND THE

Ashkun, no information of any kind is as yet available. Alt the efforts of my kind friends in Chitral and the Khaibar Pass have been unavailing. For Wai, we have some vocabularies of doubtful authority. This exhausts the list of the known languages of Kafiristan. The two about which we have any certain knowledge, Bashgali and Wast-veri, are, like Xho-war and Shina, certainly Indo-Aryan in their phonetic systems, but, also like them, possess grammars which present difficulties to the student. We know that in prehistoric times the Aryan, or Irano- Indian, language split up into two, an Iranian and an Indian. "We know also that the Iranian again split up into two branches, a Western and an Eastern. The modern representative of Western Iranian is Persian, and the most important one of Eastern Iranian is Pakshto. The modern Indo-Aryan vernaculars are the present- day representatives of the Indian branch. These throe branches are recognizable by well-defined phonetic laws. A convenient shibboleth is the Persian dast, ' a hand,' which corresponds to the Pakshto Ids and the Indian hath or hast. These four languages—Wasi-veri, Bashgali, Kho-war, and Shina — agree in following the Indian phonetic system, but in some grammatical particulars they show remarkable points of agreement with the Eastern Iranian tongues. The modern Indo-Aryan languages can all be traced back to the ancient Sanskrit spoken in Vedic times. This is true both of their phonetic systems and of their grammars, but by no course of derivation with which I am at present acquainted can I recognize the Sanskrit original* of some of the grammatical forms presented by these four. This may be my fault; it very possibly is so, for we lack the connecting link between them and the ancient language from which they are derived, which we possess in the Prakrit* for the vernaculars of India. If we had such a link, i.e. specimens of the mediaeval language spoken below the Hindu Kush, the affiliation of the four with Sanskrit might be easy, but till this is the case, the most that we can say is that while their phonetic system is the same as Downloaded from https://www.cambridge.org/core. University of Warwick, on 21 May 2018 at 04:09:30, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0035869X00027878 NORTH-WESTERN FRONTIER OF INDIA. 505

that of the Sanskrit-derived languages spoken further south, we are unable to say positively that they are derived from the Sanskrit with which we are acquainted. Judging from the well-ascertained facts regarding the origin of the modern true Indian languages, we may argue from analogy and say that it is probable that the four were derived from Sanskrit, but how they were derived, and by what stages, we are not at present able to say. This fact, together with the remarkable circumstance that some of their grammatical forms agree with those of the Eastern Iranian languages, has led certain scholars to suggest, with at least equal probability, that while the four are undoubtedly Indian, they are not necessarily Sanskritic, but are descended from a mother-dialect closely akin to Sanskrit, of which, possibly, ancient Sanskrit was a further developed form. This mother- dialect was, so to speak, left behind below the Hindu Kush, while the bulk of its speakers went on into India, and there founded the Indo-Aryan civilization, and the Indo-Aryan speech.1 While* it is impossible in the present state of our knowledge to prove this contention, it can at least be said to be not impossible, and to explain some difficult points. If it is true, then the four languages represent a stage of the Indian branch of the Aryan family older than Sanskrit, a stage which had already developed all the phonetic system of that branch, but which still retained some linguistic connection with its Iranian sisters on the other side of the Hindii Kush. It only remains to state that there is no sudden change between these languages and the definitely Indian ones of the North Panjab. The two sets merge into each other by stages. The first stage consists of the Kalasha, Gawar-bati, and Pasbai languages, about which I am now going to speak, which are almost certainly Sanskritic, yet still show remarkable points of contact with

1 An interesting point of agreement between these four languages and the Eastern Iranian ones is the infinitive in k. Thus, with the Eastern Iranian WafchI of the Pamirs, chilgak, ' to desire,' and the Orraurl of Wazlristan, (jhwik, ' to say,' compare the WasT-veri pesum-tinik, ' to beat,' the Kho-w5r bik, and the Shina boki, ' to become,' and, finally, among languages to be dealt with later on, the Kalasha hik, and the Gawar-bati and Pashai ilk, ' to be.' Downloaded from https://www.cambridge.org/core. University of Warwick, on 21 May 2018 at 04:09:30, subject to the CambridgeJ.R.A.6 Core .terms 1900 of. use, available at https://www.cambridge.org/core/terms. 34 https://doi.org/10.1017/S0035869X00027878 506 LANGUAGES SPOKEN BEYOND THE

Kho-war ; and the second of the Kohistan languages already described, which are quite certainly of Sanskrit origin. This points us to a state of affairs in the olden time which is just what might have been expected, viz. the old parent language of the four gradually merging into its sister, the Sanskrit of the north-west of the Pan jab, by insensible gradations, and not separated from it by any hard and fast lines. The Kalasha Kafirs dwell in the dbab between the Chitral and Bashgal Rivers. They have a language of their own, which is called by their tribal name. Dr. Leitner gave some information about it many years ago, which has now been supplemented by the Survey. Lower down the Chitral, at the junction with the Bashgal, in and about the country of Narsat, dwell the Gawars, who also have a language of their own known as Gawar-bati, or Gawar-speech, of which a vocabulary was given by Colonel Biddulph under the name of Narisati. Still lower down, on the right bank of the Chitral, which has now become the Kunar, dwell the Pashais, who also have a language of their own. Pashai is spoken as far west as Laghman, and extends as far north as the Waigal Yalley, though whether it is the same as the Wai Kafir already alluded to I have not yet been able to determine satisfactorily. At any rate, it is by far the most western outpost of the Indo-Aryan languages. It is an island of Indian speech in the heart of Afghanistan, and is bounded on the north by the Kafir dialects and on the other three sides by Pakshto. These three languages, Kalasha, Gawar-bati, and Pashai, are all very closely connected. They are certainly Indo-Aryan, and nearly certainly Sanskritic, though it must be pointed out that they possess some of the typical grammatical peculiarities of the four languages with which we have just been dealing.1 Kalasha, whose habitat is close to that of Kho-war, possesses most points of continuity with that language, and forms a bridge between it and the other two, which in their turn

1 For instance, the infinitive in h to which attention was called in the last footnote. Downloaded from https://www.cambridge.org/core. University of Warwick, on 21 May 2018 at 04:09:30, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0035869X00027878 NORTH-WESTERN FRONTIER OF INDIA. 507

bridge over the gap between Kalasha and the undoubted Sanskritic languages of the Panjkora, Swat, and Indus Kohistans. To complete this list of languages spoken on the north- west frontier, wandering shepherds, known as Grujars, inhabit the country between the Kunar and the eastern border of Kashmir, and perhaps still further to the east. These have a language of their own—a purely Sanskritic one—which, curiously enough, is nearly the same as the Mewarl spoken in distant Rajputana, and is closely allied to Gujarat!. Full descriptions of Kalasha and Gawar-bati appeared in the papers which I had the honour of presenting to the last Oriental Congress. Since then, through the kindness of Mr. J. G. Lorimer, I.C.S., Political Officer of the Khaibaf, I have received complete specimens of Pashai, and, as this language has hitherto been almost unknown, the following further particulars concerning it will be of interest. The only information which up to the present time has been available has been a short list of ' Pushye' * words by Burnes, and two brief vocabularies, one of LaghmanI and one of Pashai by Leech. Leech was under the impression that these two,were distinct languages, but really the names only connote two dialects of the same form of speech. Pashai, properly speaking, is the name of the language spoken by the Dehgans of Laghman and the country to the east of it. It is also called LaghmanI from the tract in which it is spoken, and Dehganl, because most of its speakers belong to the Dehgan tribe. The boundaries of the language are said to be, roughly, on the west the Laghman River, on the north the boundary of the Kafirs, on the east the Kunar River, and on the south the Kabul River, but the riverain villages on the left bank of the Kabul speak Pakshto, not Pashai. A certain number of Pakshto-speaking communities are also found interspersed

1 The only importance of this list is that the spelling of the name misled Lassen, who put it down, on Burnes' authority, as a distinct language. I owe this piece of information to the kindness of Dr. Kuhn. Downloaded from https://www.cambridge.org/core. University of Warwick, on 21 May 2018 at 04:09:30, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0035869X00027878 https://doi.org/10.1017/S0035869X00027878 to theCambridge Coretermsof use,availableat Downloaded from

LTST OF LANGUAGES SPOKEN BEYOND THE NORTH-WESTERN FRONTIER OF INDIA, en EXPLANATORY OF THE ACCOMPANYING MAP. o 00

https://www.cambridge.org/core 1 FAMILY. GROUP. LANQUAGE. DIALECT. i WHEIIE SPOKEN. 1

Scythian ? Burushaski, j Standard Hunza and Nagar. •fc-i Khajuna, or Kunjuti I Warshikwar ! Yasin and neighbourhood.

j) Tibeto-Burman Tibetan Bait! Baltistan. Iranian Western Persian Badakhshi Badakhshan. ,, Eastern I Wakhi Standard Wakhan, and near Zebak. https://www.cambridge.org/core/terms

. (Ghalchah sub-group) Shighni or Khugni i Shighnan, Eoshan, Gharan, and near Zebak. University of Warwick o Sariq-qoli Taghdumbash Pamir and Sarikol. W Ishkashami, Zebaki, or \ Ishkasham, Zebak, and valleys of the Dorah and Sangllchi ; Nuqsan Passes. MunjanT or Mungi Munjan. Yiidgha or Leotkuh- Upper part of the Lutkho Valley, south of the Hindu

, on i-war Kush. Eastern Northern or Pakhto ; Swat, Dlr, and Bajaur. Parts of Hazara, Peshawar, 21 May2018 at04:09:30 Pakshto j and the country to the west. As a lingua franca, up i ! the Indus Valley for a considerable distance. Indo- Aryan Shina-Khowar Shlna ; North-Western North-west of Gilgit. . ' GilgitI Gilgit Valley. Astori Astor Valley. BrokpaofDah-Hanu On Indus, near Baltistan, and Ladakh Frontier.

, subject Chilasi Indus Valley from near Astor to Tangir and Sazin. Kho-wSr, ArniyS, Standard Chitral and part of Yasin, Chitrarl https://doi.org/10.1017/S0035869X00027878 to theCambridge Coretermsof use,availableat Downloaded from Indo-Aryan Kafir BaHhgali Standard North Kafiriatan, especially the Bashgal Valley. WasT-veri or Veron ,, Spoken by the Presun Kafirs. Wai-gall or Wai ,, The valley of the Waigal. Aahkuu : Nothing is known about this language except its uanie,

https://www.cambridge.org/core and the fact that its speakers live to the south-west of the Presuns. Its classification is therefore pro- visional. f N.W. Sanskritic Kalasha The Doab between the Bashgal and Chitral Eivers. o Gawar-bati or Narsati ,, Round the confluence of the Bashgal and Chitral Eivers. Pashai Eastern East Laghman.

;, Western West Laghman. k fed N.W. Sanskritic Indus KohistanI or Kill Duberi Kandia, and Duber Valleys. Maiya https://www.cambridge.org/core/terms

. ?, Eastern East side of Kohistan. University of Warwick Western Western Kohistan, round Koli, Palus, Batera, Chills, I and elsewhere.

), )) ,, Southern South Kohistan.

»» Garwl Standard Swat Kohistan. fed Dir. Pi f> ) ) >j Dlrl

, on Swat and Panjkora Kohistans below Garwl. 5> )J Torwal! Standard

21 May2018 at04:09:30 Valley of Kashmir. )) t, Kashmiri

)> Chhibhali Hill country between the Kashmir Valley and the Indus Kohistan. . Central Sanskritic Hindi (KajasthanI) Gujarl Spoken by Giijars over the country to the east of the Chitral-Kunar River. i

> Between Chitral and Chilas there is a mountainous tract of country which has not yet been linguistically explored. It is said to be principally peopled by Giijars who speak their own language. There are several sub-dialects of Gujarl which differ slightly from each other. 510 LANGUAGES BEYOND N.W. FRONTIER OF INDIA.

at other places within these limits. The principal place-i and neighbourhoods in which Pashai is spoken are Barkot, Sutan, Waigal on the side next the Kafirs, Janjupur, Amlii, Sur^h,1 Badiall, Islampur, Badshali Kale, Balatak, Kunadu, Deogal, Nurgal, Chaman, Najll, Sau, Kulmiin, Tagao, Siau, and Kulab. Some of these are considerable tracts. The number of people speaking Pashai has been estimated at ] 00,000, and with regard to the Pashai region and its probable character this estimate does not appear to be unduly large or unduly small. There appear to be different dialects of Pashai, but the variations are said to be not great. The distinction drawn by the people themselves is between the ' harsh tongues ' of the hills2 and the softer tongue of the more level country. I have myself examined specimens in two dialects, a western and an eastern. The differences are mainly ones of pronunciation. Thus, the short ", which is so common in Pakshtd, also occurs in the eastern dialect, but usually appears as a long e in the western one. For instance, Eastern puthla, Western puthle, a son. Again, an Eastern sh becomes a Western k]i, as in Eastern shuring, Western khoring, a dog. In order to explain the accompanying map, I append a table giving the names of the various languages spoken beyond the North-Western Frontier of India, with their dialects and habitats.

1 The small o above the line indicates the very short a-sound known as tho fatha-e afghmil, which is so common in Pakshto. 8 One of these is called Kulmani from being spoken in Kulman.

Downloaded from https://www.cambridge.org/core. University of Warwick, on 21 May 2018 at 04:09:30, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0035869X00027878 MAP ILLUSTRATING THE LANGUAGES SPOKEN BEYOND THE NORTH-WESTERN FRONTIER OF BRITISH INDIA. Approximate boundaries only marked where known.

Index of Colour—Iranian Pamily, Blue. Indo-Aryan Pamily, Red. Non-Aryan Languages, Yellow. SCALE 32 MILES TO I INCH 70 72 73 74

ASWKUN- R / I

EASTERN